Friday 16 October 2015

The Voices of SOMA

SOMA has been a long journey. For roughly five years I’ve worked on Simon, Catherine, and the others who in some way inhabits the world of SOMA. The game has changed a lot over the years, not just in gameplay and plot, but in tone as well. I think one of the most grounding and solidifying aspects of creating the characters for SOMA has been giving them an actual actor’s voice.

Not exactly Fallout amount of dialog, but still pretty good

2013 - A Vertical Slice
Voices were first introduced as we were completing the “Vertical Slice” in February 2013. It was to be our first self-contained build that we could show off, test player reactions, and to allow ourselves to make a better judgment on our own work. For me this meant that I would head off to SIDE studios in London and record all the voices needed for about ten levels or so. At the time of the first recording, everything was still very much up in the air. I had written lines for the monsters inside Curie – even the WAU! It was just a mess and I’m glad we cut most of it out.
We haven’t really kept much at all of the material that we recorded during these sessions, but it did raise a very important question – what should Simon and Catherine sound like?

Simon and Catherine are by far the biggest and most complex roles Frictional Games have ever done. Neither Simon or Catherine fill any particular stereotypes found in games, which meant finding the right actors wasn’t going to be easy.

Our initial thought for Catherine was to use her background from Taipei and have her talk with a noticeable accent. In my book, accents are great. Not just because it’s nice to be diverse and inclusive, but because it distinguishes a character and gives them so much flavor. It’s very useful, especially for a game like SOMA where most of the characters don’t have visibly human features to make them stand out. However, what we found after trying a few accents on Catherine was that it actually didn’t work the way we wanted. Combining the voice with the non-expressive visuals made it a little cartoonish, reminding people of Amy Wong from Futurama, a character who is largely comedic. Catherine at this point was even more introverted and it didn’t fit at all that the player should find her amusing like that. Looking back I think we could have made it work, but that Catherine would have been a very different Catherine from what we ended up with. To minimize the risks of making an overtly funny character we decided to let the actress almost completely drop Catherine’s flavorful accent and instead move towards British RP.

Simon’s voice kind of had an opposite journey to Catherine as we ended up finding Jared Zeus right away. However, it wasn’t going to be quite as simple as that.

Jared had taped microphones on his head so it would sound awesome
when playing with headphones - appreciate it!

2014 - The GDC Build
The next recording was almost exactly a year later. This time we were looking at the first version of the game that we could show journalists, mainly at GDC, the game developer conference in San Francisco that same year. This was basically the same part of the game we recorded the year before, but with substantial rewrites, which meant going back to page one and doing it all over again.

A big change from my perspective was that it had been decided that we should try to find a tougher kind of Simon, and so we casted another actor to play that part. It caused some slight weirdness seeing how at the same time I had sort of gotten into the Jared Zeus rhythm and started to write him with that actor in mind. During the last year Simon, along with Catherine, had also become more relaxed. I wanted to adopt a little more comical absurdity than just pure angst. Not to go too deep into style, but basically the changes was meant to move away from the melodrama that is more effective when considering a character like Daniel in Amnesia and try to approach a style of horror with a little levity which would be more forgiving when doing proper conversations. Going with a tougher Simon kind of sidetracked that everyman feeling that I had written and this turned out to be a case of wronging a right.

While we were dealing with Simon’s voice, we still didn’t have a Catherine that we felt comfortable with either. The actress we had was great, but the voice simply didn’t gel with the character we wanted. As a side-note I would say that this is how most of these things work out. When choosing an actor for a part there’s very rarely any question of who is better or worse, it’s much more about the personifying qualities that the specific actor can provide. Auditions are not so much a proof of acting chops as much as it is a question of can we find an interesting take on this character that would serve the story. So it’s not the best actor, but the best suited one that you want.
Anyway, back to casting Catherine. All we knew was that the voice we had didn’t work the way we wanted. We did a regular casting round, but none of the voices really hit home – probably because we didn’t really know what to ask for either.
This was starting to look like a serious problem, we were missing one of the most important characters of the game. I went over all the auditions for Catherine again and again, and then I started to look outside of Catherine and dug into all the other roles we had been casting. That’s when I stumbled on Nell Mooney. Nell had been reading for the role of Alice Koster and there was just something about that voice that made sense to me. She had a warm, likable voice that made me think: if this person says or does something completely insane I might just still forgive and trust her. Which is kind of exactly what Catherine needed to be. Thankfully, Thomas agreed that we should give Nell a shot and we ended up casting her even though she hadn't read any of Catherine's lines. A really lucky break for us considering how incredibly well Nell would come to bring that character to life.

Chuck helps people getting into action mode!

2014 - Building on What We Got
The next recording in October 2014 was meant to complement the stuff we already had. Unlike the jump between the first two recordings, this time it was important that we could get the same actors back to continue their roles. This is the first time we found some logistical problems with recording like this. Unsurprisingly, actors are people that do a lot of cool stuff and they can’t just sit around waiting for us to call them up again. It was really crushing to replace some awesome performances due to the fact that we couldn’t get the actors back that we needed. This meant that we would need to replace a lot of actors and record the old stuff again as well.

Since the last recording it had become apparent that the tougher Simon wasn’t really working the way we wanted, so it was decided that we were going make another round of auditions for that role. Thankfully, I was able to convince Thomas to go back Jared Zeus again.
Even though some of the minor roles still needed to be re-cast and re-recorded, we had our two main characters Simon and Catherine sorted out, which was a huge load off my mind.

Preparing a session of Simon (Jared) talking to Lindwall (Andrea Deck)

2015 - Beta and Pick-ups
I think it is safe to say that the time between October 2014 and February 2015 is when the game finally found its form. A tremendous amount of stuff had been rewritten and in a way this is when we recorded the game. It was a two week recording and it was a blast – probably the most fun I’ve had professionally. Every morning Jared and Nell would come in and record for four hours. They had by far the most lines and there was a lot to cover. After lunch we would continue with some of the smaller roles, since it’s not a good idea to kill the voice of an actor by keeping them for a full day. Having Jared and Nell consistently record every morning turned out to be great idea. I could kind of hear the whole game play out chronologically and I could see them react to my explanations of what was happening in the different scenes in a way that a player would pick up things. A game script isn’t very descriptive, so even though they most likely read through the lines before, it’s not really until I give them the context that they realize what is going on. So I could watch them react to the story as I was telling it. Kind of reminds me hosting pen-and-paper RPGs. And the great part about that was being able to do minor changes to the script so it fell more closely in line with what the actors were expressing. I think this is the time when I finally consciously started to express some of the emotional themes of the game, like hope and denial. It was when talking to Damien, our amazing director, that enabled me to slow down and not think like a content producer and instead discuss and work through the material in a sensible way.
We ended up doing one final round of recording late spring where we filled in some gaps in the story and did a new take on the enemies, but other than that I consider the February recording our definitive and by far the most important recording of the whole game. 

Wine helps Simon and Catherine cope with the horrors of SOMA

Closing thoughts 
I’m not completely sure why I wanted to write an article about the voices of SOMA – I didn't even tell you about all the fun stuff like how Akers' once had a voice as a monster and how we almost flew an actress from Iceland to record for an hour. I like to think this is me just paying a small tribute to all the amazing work I’ve seen from directors, sound engineers, producers, casting agents, and of course the multitude of actors who at some point acted out the lines that I put in front of them. Voice recording is an incredibly fun world to work with and I hope to see all of you again in future projects.

Monday 5 October 2015

Hiring: Producer / Project Manager

Frictional is looking for a producer / project manager to employ full time.

First up, the two most basic requirements:
  • That you live in Sweden (preferably in Skåne) or can move here.
  • That you speak Swedish.
På grund av detta kommer resten av ansökan att vara på svenska.

Som producer / project manager på Frictional Games kommer du att få väldigt varierande uppgifter. De huvudsakliga sysslorna kommer att vara följande:
  • Sköta stor del av företagets planering och budget.
  • Att överse och förbättra företagets arbetsmetodik.
  • Göra planer för PR och sköta kontakter.
  • Ansvara för kontakten med butiker, förläggare, outsourcers och andra samarbetspartners. 
  • Följa upp och utvärdera diverse förslag från andra företag.
Egenskaper vi vill att du ska ha är:
  • Beredd på att jobba hemifrån. Frictional Games har inget kontor.
  • Tidigare erfarenhet att driva eller vara manager på ett IT-företag. Bakgrund inom spel är ett plus, men inte nödvändigt.
  • Inte rädd för att jobba med kalkylark och liknande.
  • God datorvana.
  • Förstår dig på grundläggande ekonomi.
  • Goda kunskaper i personaldynamik.
  • Självständig, inte rädd för att ta ansvar och god initativförmåga.
  • Inte rädd för att lära dig nya saker.
  • Tala flytande engelska.
Tycker du att detta låter som något för dig, skicka ett mejl till

Thursday 1 October 2015

SOMA - 10 days after launch

SOMA has now been out in the wild for 10 day so it felt fitting to write a summary of how things have gone so far. But first a little trailer:

I'm going to start with what I think most people are interested in: how much has the game sold? The current number now is at about 92,000 copies across all platforms (due to legal reasons we can't give a per-platform breakdown). This is quite good for 10 days (+ preorder time) of sales! The money that we've got from this will pretty much pay our company expenses for another 2 years. Sales are still going pretty strongly too, with a total of around 2,000 copies sold per day. This number is bound to drop over time, and it'll be interesting to see just how fast and where it stabilizes. While a lot of sales obviously come close to launch, a big part of our normal earnings comes from a slow daily trickle over the years of our existing titles. So our average daily sales a month or so from now on is actually more important than all of the units sold up to this point.

How does this compare to our other releases? Well, Amnesia: The Dark Descent sold 30,000 copies in the first month (and around 20,000 the first week). So SOMA's launch is obviously a lot better than that. Compared to Amnesia: A Machine For Pigs, though, the launch is a little bit worse. That game sold about 120,000 copies the first week.

Our goal for SOMA's sales is 100,000 after a month, and at the current pace it should be able to reach pretty much exactly that with a few units to spare. However, this doesn't mean that we've come close to recouping all our costs. We need to sell almost 3 times that amount to do that. But given that it took us 5 years to make the project, there's no immediate stress to do so. One of the great things about funding SOMA 100% ourselves is that all money earned goes into our own pockets and is directly used to fund our upcoming projects. So we are under no pressure to recoup immediately so long as we get enough to keep going - which we certainly have now.

Finally, another very interesting aspect is how new titles tend to cannibalize on the previous ones. We saw this with A Machine for Pigs; after it launched the daily sales of The Dark Descent were almost cut in half. That was not that unexpected though, given that they are both from the same franchise, but still a bit weird that the games' combined sales ended up being pretty much what The Dark Descent sold on its own before. What we didn't expect was for SOMA to do the same. When the pre-orders for SOMA started, Amnesia sales dropped by about 30% or so and this drop still remains. This feels strange as the two games are not connected apart from being made by the same company, so we wonder what mechanism it is that causes this. It might be that Amnesia's sales will rise again a bit later on though, so it's too soon to tell yet just what the effects are.

The critical reception of SOMA has been, overall, really, really great. MetaCritic is currently at 85 and the Steam reviews are 94% positive.

The thing that I worried most about personally was how the themes would be received. It turns out that I needn't have worried - that's the thing we have fewest problems with. Even reviews that gave us so-so scores lauded the game for the thought-provoking narrative. This feels awesome, as this has been the core focus during our five years of development.

The most common issue people have had is that they've felt the game wasn't scary enough. This is quite interesting, so I'd like to take a little time to discuss this.

One reason this was so is probably due to expectations. While we've tried to be very clear that SOMA will be a different game from Amnesia: The Dark Descent, we have still used the name "Amnesia" as a way to grab attention. This sends a bit of a mixed message, as people might simply assume that because we say "from the creators of Amnesia", a similar experience will be provided. One idea would have been not to mention the studio's heritage, but that feels stupid from a PR perspective. Another idea would have been to tone it down a bit, but it's hard to say exactly how to do that. The fact of the matter is that SOMA, just like Amnesia,  is very much a horror game. It's just that it is presented in a different manner, using slower build-up and more focus on the psychological aspects.

Another reason why some people felt it was not scary enough is because horror is extremely subjective. The reactions to how scary SOMA is range from "not at all" to "the scariest game I have played", and some of the people in the latter camp are survival horror veterans. We had this sort of reaction to Amnesia: TDD as well, but it feels even more spread out for SOMA. When we released The Dark Descent, horror with no combat was still a very fresh concept, but five years later that is no longer the case, and it has lost its impact for some people. SOMA also employs a riskier approach to monster AI that assumes the player will act in certain ways and reach a certain understanding about how the creatures work. If players don't do this the experience might suffer. Above all, the main horror in SOMA is supposed to come from the existential dread that's slowly unveiled as the game progresses. And in order for this to work properly, a lot of pieces need to align, and it will not work for everyone.

So in the light of that, it doesn't feel all that bad that we didn't get a more universal praise for the game's scariness. But it's taught us a valuable lesson: that one should be very careful in managing people's expectations. This is a lesson that we thought we knew after A Machine For Pigs (which didn't turn out to be the game many wanted it to be) but apparently we hadn't learned enough. Once your studio gets associated with a particular game, it'll play a huge role in what people expect from upcoming releases. That said, the vast majority of people that had expected another Amnesia ended up enjoying SOMA once they realized the game was different. So I don't feel it has been a complete failure by any means, but just one of those things that needs more work in the future.

It is interesting that this is no longer a subject brought up much. So I thought I would quickly get into it. And the first thing to note is that SOMA is the first game we have launched without having a pirated version out before release!

Another thing I have noticed is that we get fewer tech support requests from people with pirated versions than we used to have. It's often pretty easy to spot these people as we issue new patches frequently, so there are lots of telltale signs in the log files. I'm not sure if this means piracy has decreased for SOMA, or if these people find tech support elsewhere, but I felt it worth mentioning.

As for us personally, we haven't even talked about piracy. The only time it matters to us is when sending out review copies. Amnesia had a pirated version leaked before release, so now we make sure that we at least send out a DRM-protected version of the game to reviewers. But other than that, I don't think we've discussed it for even a second. This is quite different from back in 2007 when I know me and Jens had hours of discussions on the subject.

I've already touched upon this above when discussing the game's reception. However, how to market SOMA in terms of horror was the easy part. The hard part was to explain what makes the game special. When we released Amnesia, showing off the physics and explaining that you couldn't fight back was more than a enough for the game to stand out. But now the market is filled with these types of games, and more is needed to get people excited.

The main unique feature of SOMA is its exploration of consciousness and what it means to be human. This is also what has been the most celebrated feature of the game after launch. But explaining this to press and gamers prior to release has been exceptionally difficult. This is not some gameplay gimmick that can be shown off during a short demo session, but something that requires hours of build-up. So when you talk about the game, you have to be fuzzy and talk about very high-level concepts. When doing interviews like this I often got the impression that I wasn't really taken seriously. The press don't expect any lofty design aspirations to come true and would rather hear about concrete and more easily-digested (and explained) features.

To make things even harder, SOMA is very hard to talk about without spoiling the experience. I could never give an example of exactly how we handle our thematics through gameplay without spoiling a big chunk of the game. This problem of spoilers also makes the game hard to demo and to give to YouTubers. If we just give people a part of the game where you are chased by monsters, that would misrepresent the game (making the expectation problem worse) and fail to explain what is so special about SOMA. And if we show off one of slower sections that are all about build-up, mood and thematics, we have to show off really long segments, which becomes too spoiler-filled and takes way too much time for a demo. (For more discussion on making a demo for SOMA, see here).

We solved the YouTuber issue by only sending it out to a few trusted people, and only allowing a maximum of 15 minutes to be shown. That way we got people to play a lengthy part of the game (around 3 hours) and then show a distilled, and fairly spoiler-free, video to their viewers. We could only do this pretty late in development though, and given how important streamers and YouTubers are for PR these days, it felt like we would have like to do more earlier.

Another issue is that we might have unveiled the game a bit too early. We first showed off SOMA back in October 2013 and the plan was to keep content coming out until release. This turned out way harder to keep up with than what we'd initially thought. Because we were so unwilling to spoil the game, we could provide very little in terms of playable material for the press. Because of this, we had issues getting proper coverage at the end, as most of the standard things like "first playable preview" had already been done over a year back. We'd also had a plan to release a monthly live-action video clip to keep interest up, but because of production problems it got delayed and this plan fell through. (We are however showing them now!)

So it feels like it might have been better to have unveiled the game a year or so later to be able to keep up interest all the way to release and to have a more massive promotion campaign that way. A big issue with that is that it would have been very bad for the team morale. It's quite hard to work on a project in the dark for several years, and there was a very evident boost in spirit once we had let the world know that SOMA was coming. Added to this is that we got a lot of good feedback from press and fan reactions, which helped us shaped not just our PR but the actual game too. This is makes it much more uncertain if a later unveiling really would have been a better move.

So what is next for Frictional Games? First of all, now just about all of the major post-release issues have been patched up, most of the team will take some rest. We'll then focus a bit on documenting how the game and engine works, in the hopes that modding will reach the glorious heights it did for Amnesia. After that we are on to new secret projects.  But those secret projects are really secret, so we can't say a word.

Finally a gigantic thanks to all who have bought the game! We love hearing about your experiences so please tweet, comment on Facebook, or leave a comment here and say what you thought about the game!